An Analysis of Industrial SRAM Test Results - A Comprehensive Study on Effectiveness and Classification of March Test Algorithms
Modern microcontroller devices contain large numbers of embedded SRAM memories. Especially microcontrollers for automotive applications are highly safety critical and so are their embedded memories. Zero defect density of memories is an essential safety requirement and defect coverage of memory tests is of special interest during production, memory testing and burn-in. A comprehensive study on memory fault models and test algorithms was performed during full industrial production on embedded memories of automotive microcontroller devices to obtain statistical data on memory faults and their correlation to test algorithms. Due to the large quantity of devices, a statistical resolution of parts per million could be achieved. A detailed analysis of the resulting data shows the efficiency and similarity of memory test algorithms could be analyzed. The relation of test algorithms to many specific faults or sets of faults was used to classify algorithms and faults. This classification serves as a foundation for an efficient test selection.
Classification, SRAM test, March Algorithms, Functional Fault Models
Klaus Oberlander, Ulf Schlichtmann, Alfred Eder, Michael Linder, "An Analysis of Industrial SRAM Test Results - A Comprehensive Study on Effectiveness and Classification of March Test Algorithms", IEEE Design & Test of Computers, , no. 1, pp. 1, PrePrints PrePrints, doi:10.1109/MDAT.2013.2279752